A tax, we naturally think, is a transfer: the state takes a sum from the taxpayer and adds it to the treasury, and what one loses the other gains, the books balancing. This is true as far as it goes, but it misses the deepest fact about taxation, the one Jules Dupuit was first to see: a tax destroys more than it transfers. Besides the revenue it collects, a tax kills trades that would otherwise have happened — mutually beneficial exchanges that the tax makes unprofitable — and the value of those lost trades is a pure loss, collected by no one. The tax is not merely a transfer but a wedge, and the wedge does damage.
Picture any market in equilibrium: buyers and sellers meeting at a price, every trade that benefits both parties getting made. Now impose a tax on each unit sold. The tax drives a wedge between the price the buyer pays and the price the seller receives: the buyer now pays the seller's price plus the tax, or, what is the same, the seller now receives the buyer's price minus the tax. The two prices, once equal, are now separated by the tax — the buyer pays more, the seller gets less, and the gap between them goes to the state. This wedge is the essence of a tax: not a single price paid by the buyer to the seller, but two prices, separated by the state's claim, with the buyer's price above and the seller's below the old equilibrium.1
The wedge does more than separate the prices; it shrinks the quantity traded, and there lies the damage. At the higher price they now face, some buyers withdraw; at the lower price they now receive, some sellers withdraw; and so the quantity traded falls below what it was. The trades that no longer happen are the casualties of the tax: exchanges that would have benefited both buyer and seller — the buyer valuing the good above what the seller would have accepted — but which the tax wedge has made unprofitable, because the buyer's value, though above the seller's cost, is not above the seller's cost plus the tax. These lost trades yield no revenue to the state, for they do not happen and nothing is collected on them; and yet they represent real gains from trade forgone — value that would have been created and now is not. This forgone value is the deadweight loss, or excess burden, of the tax: the loss to society over and above the revenue transferred. The tax costs the private economy not just the revenue it pays to the state, but this further, uncollected loss — the gains from the trades the tax killed.2
Jules Dupuit, a French engineer studying the worth of public works, was the first to measure this loss, and the first to see that the cost of a tax is not the revenue it raises. Analyzing the toll on a bridge — itself a kind of tax on crossing — he saw that the toll's cost to society was not the money collected, which merely passed from travelers to the bridge's owner, but the value of the crossings the toll prevented: the travelers who, faced with the toll, did not cross, though crossing would have been worth more to them than the toll's real cost. That forgone value — the lost benefit of the prevented crossings — was the toll's true social cost, and it appeared in no ledger, for it was the value of things that did not happen. Dupuit captured it as a triangle: the area between the demand curve and the supply curve, over the range of trades the tax destroyed — the lost surplus that later economists would call the deadweight loss. The insight founds the economics of taxation: to count the cost of a tax, count not the revenue but the revenue plus the triangle, the trades transferred plus the trades destroyed.3
The deadweight loss has a feature that governs all of tax policy: it grows with the square of the tax rate. Double the tax, and the wedge doubles, but the loss more than doubles — it roughly quadruples; treble the tax, and the loss grows ninefold. The reason is geometric: the lost-surplus triangle has both a height, the tax, and a base, the reduction in quantity, and both grow with the tax, so the triangle's area grows with the tax squared. A small tax destroys few trades and little surplus; a large tax destroys many, and disproportionately. From this single fact flow the central maxims of tax design. Spread the burden: many small taxes on broad bases do less damage than one large tax on a narrow base, because the loss grows with the square of the rate, so lower rates over wider bases minimize the total triangle. And tax lightly what is heavily traded, for the larger the market, the larger the base over which the squared loss compounds. The square rule is why economists prefer broad, low taxes to narrow, high ones — not from a love of uniformity, but because the deadweight loss is convex in the rate, punishing concentration.4
The wedge, then, is the structure of a tax and the deadweight loss its hidden cost: the tax destroys trades, and the value of those trades is a loss no one collects, growing with the square of the rate. But the wedge raises at once the question the volume is built around. The wedge separates the buyer's price from the seller's, the buyer paying more and the seller receiving less; but how much more does the buyer pay, and how much less does the seller receive? The wedge has a total size — the tax — but how is that total divided between the two? Does the buyer bear most of it, in a higher price, or the seller, in a lower one? This is the question of incidence — of who actually bears the tax — and it is, the author tells us, the widest gap in all of policy between intention and effect. For the law that levies the tax names who shall hand it over; but the division of the wedge, the real bearing of the burden, is decided not by the law but by the market. The next plate turns to that division: how the wedge is shared, and who, in the end, pays.5
The wedge is what a tax is: not a mere transfer from taxpayer to treasury, but a separation of the buyer's price from the seller's, which shrinks the quantity traded and destroys the gains from the trades that no longer happen. Dupuit was first to see that this destruction — the lost surplus, the deadweight loss — is the true cost of a tax, over and above the revenue it raises; and that the loss grows with the square of the rate, so that broad, light taxes do less damage than narrow, heavy ones. The tax costs the economy the revenue plus the triangle. But the wedge, having a size, must be divided between the buyer and the seller — and how it is divided, the real incidence of the tax, is not what the law decides but what the market does. To that division, and to the gap between who the law names and who the market burdens, the volume now turns. The next plate is the shift.